Skip to content

Crime and Local Inequality in South Africa

Literature › Works

Gabriel Demombynes; Berk Ozler (2002). Crime and Local Inequality in South Africa. World Bank Policy Research Working Paper 2925. The World Bank, Development Research Group.
Type working paper
Scope Post-1994 South African economy
Method econometric
Period studied 1996–1996
Regions South Africa (all nine provinces), 1,066 police station jurisdictions
Themes inequality poverty unemployment, methods data sources, comparative theory, demography urbanisation, transition post apartheid
Citation key demombynes2002b

Argument

Using 1996 crime counts, census-based small-area welfare estimates and negative binomial regressions, the paper tests whether local inequality explains crime across South African police station jurisdictions. Property crime, especially residential burglary and vehicle theft, tracks mean expenditure (a proxy for returns to crime). Burglary is also higher in the locally wealthiest jurisdiction, which fits the idea that burglars travel. Inequality stays positively correlated with burglary and violent crime after controls, which the authors read as support for sociological theories. Decomposition by race shows the association comes almost entirely from inequality within racial groups, not between them. The authors treat the results as correlations and stress the limits of cross-sectional data.

Evidence and method

1996 Population Census matched to police station jurisdictions; 1996 South African Police Service crime data for 1,066 jurisdictions; per capita expenditure and inequality (mean log deviation) imputed by small-area estimation from the 1996 Census, the 1995 October Household Survey and the Income and Expenditure Survey; the 1998 Victims of Crime Survey (3,899 observations) used to correct for underreporting of burglary.

Findings

  • In the basic specification, inequality is positively associated with all four crimes, with elasticities ranging from 0.53 to 1.17. (p. 21)

    The elasticities range from 0.53 for rape to 1.17 for burglary.

  • Once mean expenditure is controlled for, the inequality-vehicle theft correlation disappears. Mean expenditure is the strongest correlate of property crime. (p. 21)

    Once mean expenditure is controlled for, the correlation between inequality and vehicle theft completely disappears

  • Burglary is higher in the wealthiest jurisdiction of a criminal catchment area, consistent with burglars travelling to high-return areas. (p. 22)

    burglary rates are 20% higher in jurisdictions that are the wealthiest among their neighbors. This is consistent with the hypothesis that burglars travel to neighboring areas where the expected returns are highest.

  • Between-race differences account for a substantial share of inequality within a jurisdiction on average. (p. 23)

    on average, 36% of inequality in a police station jurisdiction can be attributed to differences in mean levels of expenditure between racial groups.

  • The crime-inequality association is almost entirely a within-racial-group phenomenon. (p. 23)

    Almost all of the association between inequality and levels of crime can be attributed to inequality within racial groups.

  • Violent crime is related to inequality among Africans, Whites or others, not to inequality across the groups. (p. 23)

    It is inequality among Africans, Whites, or others, and not inequality across these groups that is most related to violent crime levels.

  • Underreporting varies by crime type: it is worst for robbery and least serious for vehicle theft. (p. 24)

    According to the VCS data, underreporting is a particularly serious problem for robbery and is least serious for theft of vehicles.

  • Correcting burglary counts for underreporting attenuates the inequality and expenditure elasticities and makes unemployment positive and significant, so misreporting can bias estimates. (p. 25)

    As underreporting is more likely in areas with higher unemployment, the adjustment causes the elasticity of unemployment to also become positive and significant

  • The results contradict the claim that ascribed (racial) inequalities are particularly criminogenic. (p. 30)

    This finding is at odds with the suggestion that "...economic inequalities matter, but ascribed inequalities do so particularly".

  • Poverty could not be separately examined because it is almost collinear with mean expenditure in the data. (p. 31)

    mean expenditure and poverty are very highly and negatively correlated in our data set - the correlation coefficient is -0.87.

  • The authors caution against reading the results causally. (p. 31)

    We hesitate to draw policy implications from analysis using cross-sectional data without a strong identification strategy